As many of you will be aware, one of the key industry trends in recent times has been the emergence of Service Oriented Architecture (SOA) and Service Oriented Business Applications (SOBA).
With this in mind, I'm curious to know if anyone is considering moving to SOA, whether anyone has already made the move, started to make the move, or if not, why. For those who have moved or at least started, what was/is your experience. For those who are not considering it, what are the reasons, is there no business benefit, does it look technically too much work for not enough gain? Is the concept of n-Tier and AppServer enough to cope with for now !!
Why my curiosity, well one obvious topic area for OpenEdge Principles is SOA & SOBA. Is it an area we should look to invest some time into providing guidance and samples, or do you feel there are far more pressing areas where you feel we should concentrate our efforts?
I'll be interested to know your thoughts and ideas on SOA, and if not SOA where you feel we should concentrate our efforts.
Thanks
Mike
In my case, SOA/SOBA isn't that important since I generally work with legacy code systems in the SMB space, and they don't have 'that' much interaction with other systems. The various licensing issues associated with trying to implement SOA are problematic as well.
I'd be curious though, for all the noise associated with it, how much penetration has SOA actually gotten?
In my case, SOA/SOBA isn't that important since I
generally work with legacy code systems in the SMB
space, and they don't have 'that' much interaction
with other systems.
And I guess there is no thought to break the app up to be more service based within application modules?
The various licensing issues
associated with trying to implement SOA are
problematic as well.
In particular?
I'd be curious though, for all the noise associated
with it, how much penetration has SOA actually gotten?
Me too
And I guess there is no thought to break the app up to be more service based within application modules?
Based on what business case?
The various licensing issues associated with trying to implement SOA are problematic as well.
In particular?
I'll leave that as an exercise for you. Take a new OE 10 system w/a webservice which talks to an unknown number of client and non-client-owned systems, and tell us what the licensing cost is for just that service.
PSC license structure "right now" is based named user / registered client, and web services don't really fit into either model.
And I guess there is no thought to break the app
up to be more service based within application
modules?
Based on what business case?
Although I probably didn't put it that way, that's what I was trying to alluding to, that there is no obvious business case to justify making the changes.
The various licensing issues associated with
trying to implement SOA are problematic as well.
In particular?
I'll leave that as an exercise for you. Take a new OE
10 system w/a webservice which talks to an unknown
number of client and non-client-owned systems, and
tell us what the licensing cost is _for just that
service_.
PSC license structure "right now" is based named user
/ registered client, and web services don't really
fit into either model.
Ok, so I walked right into that one Not being in Product Management I can't really comment on possible licensing changes, but such a topic was discussed on the PEG (http://www.peg.com/lists/peg/history/200606/msg00188.html) !
Ok, so I walked right into that one Not being in Product Management I can't really comment on possible licensing changes, but such a topic was discussed on the PEG
Understood, but that's sometime in the undefined future. Right now with the current licensing structure, SOA is hard to fit in the SMB space I play in.
Ok, so I walked right into that one Not being
in Product Management I can't really comment on
possible licensing changes, but such a topic was
discussed on the PEG
Understood, but that's sometime in the undefined
future. Right now with the current licensing
structure, SOA is hard to fit in the SMB space
I play in.
Touche
At Exchange 2006 in Las Vegas we made a preliminary announcement that we will be introducing very soon Access Agent pricing/licensing for OpenEdge 10. The model is designed to provide a cost-effective licensing solution to applications with "occasional" or "unknown" users/clients.
You will be hearing more about it through regular channels, and/or contacting your account manager.
While I am sure that lots of us will be glad when real-soon-now arrives, I'd like to flip back to the original question to say that I think that SOA is extremely important as a technique for evolving existing applications ... once we get past these licensing hassles. While I am working on techniques for wholesale upgrade of entire applications, the next best thing has got to be bringing in a bus and progressively moving individual component onto it as needs arise. It is a lovely way to clean up a legacy mess without ever having take any big leaps.
..... While I am working on techniques
for wholesale upgrade of entire applications, the
next best thing has got to be bringing in a bus and
progressively moving individual component onto it as
needs arise. It is a lovely way to clean up a legacy
mess without ever having take any big leaps.
So you see a services based architecture as a good way to approach transformation, or are you saying that if your going to transform, why not move to a services based architecture?
Assuming a services based app is the target, are you thinking that everything is converted to a service, including internal communications between components on what one might term a 'micro-bus', or just those services that would be externally available via an ESB?
Hi.
I may be a little late , but I would like to share some of my thoughts about SOA.
We have here in Latin America basically two types of partners: theb ones who are really interested in bringing their applications to a SOA model and other ones totally immersed in the SMB somewhat narrow vision and more worried about solving short-term business pains.
Of the first kind we have the likes of Datasul and Dual Software. The first is an industry giant, which is really going towards a new concept for their apps. They've developed a wholly new programming framework that, among other features, allow them to produce new services-based software. While the main purpose of this approach for the nearest future is to enable their applications to expose their logic through an Adobe Flex/Flash presentation infrastructure, the so-developed applications are ready to go in a SOA environment.
Dual Software is a small partner whose main product is a healthcare-oriented ERP. They are walking the first steps towards SOA, using the same Datasul's "presentation-first-integration-later" service approach, and are really evolving with an Ajax-based front-end.
On the other side, as I said, there are market niches that don't see advantage in moving towards SOA at all.
I would like to keep in touch to know how this point has been addressed in different parts of the world.
Thanks,
Alessandro
So you see a services based architecture as a good way to approach
transformation, or are you saying that if your going to transform, why not
move to a services based architecture?
Both, actually. To me, SOA is a universal goal to which we should all be aspiring, regardless of whether we are looking at small changes or wholesale architectural reform. I have no question that the target of a wholesale transformation should be ESB/SOA, but I also think that SOA is a good incremental transform strategy as well.